A data product is a reusable, self-contained package that combines data, metadata, semantics and templates to support diverse business use cases. It can include components such as datasets, dashboards, reports, machine learning models, pre-built queries or data pipelines.

The concept of data products gained prominence in 2019 when Zhamak Dehghani introduced data products as a core component of the data mesh architecture.¹

Data as a product introduces a new unit of logical architecture called data quantum, controlling and encapsulating all the structural components needed to share data as a product - data, metadata, code, policy, and declaration of infrastructure dependencies - autonomously.²

Characteristics

Data as a Product adheres to a set of usability characteristics²:

1. Discoverable
2. Adressable
3. Understandable
4. Trustworthy and truthful
5. Natively accessible
6. Interoperable and composable
7. Valuable on its own
8. Secure

Marty Cagan's characteristics of successful products²:

• Feasible
• Valuable
• Usable

Types

Databricks training material² distinguishes following types of Data Products:

• Source-aligned Data Product - usable and relevant representation of source data. This is private data asset that are not shared with others.
• Derived Data Product or Data Product - cleansed and enriched data asset designed for analytical usecases. It provides a single source of truth with a unified view across the domain (or subject area) and consistent data definitions. This type of data asset is shared, reusable and is available across the organization.
• Customer-aligned Data Product - derivative type of Data Product, built on lower-level Data Products. It is designed for specific purpose for end-user(s) - e.g.: dashboards, reports, calculations. May or may not be shared across the organization.

---

1. What is a data product? (2026) IBM. ↩︎

2. Dehghani, Zhamak (2022) Data Mesh. Delivering data-driven value at scale. O'Reilly Media, Inc. ISBN: 978-1-492-09239-1. ↩︎